Active perception and disentangled representations allow continual, episodic zero and few-shot learning

2026-02-22

AI Summary

This work addresses the challenge in conventional continual and few-shot learning methods, where entangled representations are prone to catastrophic interference, hindering the simultaneous achievement of rapid adaptation and strong generalization. To overcome this, the authors propose a Complementary Learning System (CLS) that disentangles representations by implementing a fast learner as a context-driven episodic memory module—used not merely for replay but to generate contextual biases. These biases guide a slower statistical learner to encode novel stimuli in a structured manner. Integrated with an active perception mechanism, the architecture enables continual learning in both zero-shot and few-shot settings. Experiments demonstrate that the approach effectively mitigates representational interference and exhibits robust rapid learning and generalization capabilities under observational variability and uncertainty.

Technology Category

• Machine Learning: Life-Long and Continual Learning
• Computer Vision: Representation Learning for Vision
• Multiagent Systems: Multiagent Learning

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization

Abstract

Generalization is often regarded as an essential property of machine learning systems. However, perhaps not every component of a system needs to generalize. Training models for generalization typically produces entangled representations at the boundaries of entities or classes, which can lead to destructive interference when rapid, high-magnitude updates are required for continual or few-shot learning. Techniques for fast learning with non-interfering representations exist, but they generally fail to generalize. Here, we describe a Complementary Learning System (CLS) in which the fast learner entirely foregoes generalization in exchange for continual zero-shot and few-shot learning. Unlike most CLS approaches, which use episodic memory primarily for replay and consolidation, our fast, disentangled learner operates as a parallel reasoning system. The fast learner can overcome observation variability and uncertainty by leveraging a conventional slow, statistical learner within an active perception system: A contextual bias provided by the fast learner induces the slow learner to encode novel stimuli in familiar, generalized terms, enabling zero-shot and few-shot learning. This architecture demonstrates that fast, context-driven reasoning can coexist with slow, structured generalization, providing a pathway for robust continual learning.

Problem

• continual learning
• few-shot learning
• disentangled representations
• catastrophic interference
• complementary learning systems

Innovation

• complementary learning system
• disentangled representations
• active perception
• continual learning
• zero-shot learning